Therapeutic activation of endothelial sphingosine-1-phosphate receptor 1 by chaperone-bound S1P suppresses proliferative retinal neovascularization

Abstract

Sphingosine-1-phosphate (S1P), the circulating HDL-bound lipid mediator that acts via S1P receptors (S1PR), is required for normal vascular development. The role of this signaling axis in vascular retinopathies is unclear. Here, we show in a mouse model of oxygen-induced retinopathy (OIR) that endothelial overexpression of S1pr1 suppresses while endothelial knockout of S1pr1 worsens neovascular tuft formation. Furthermore, neovascular tufts are increased in Apom^-/- mice which lack HDL-bound S1P while they are suppressed in Apom^TG mice which have more circulating HDL-S1P. These results suggest that circulating HDL-S1P activation of endothelial S1PR1 suppresses neovascular pathology in OIR. Additionally, systemic administration of ApoM-Fc-bound S1P or a small-molecule Gi-biased S1PR1 agonist suppressed neovascular tuft formation. Circulating HDL-S1P activation of endothelial S1PR1 may be a key protective mechanism to guard against neovascular retinopathies that occur not only in premature infants but also in diabetic patients and aging people.

Keywords: Sphingosine-1-phosphate (S1P); neovascularization; vascular retinopathy.

Figure EV1. S1PR1 signaling in retinal EC inhibits neovascularization in OIR

1. Schematic representation of the breeding strategy to establish S1pr1 ECTG: females carrying the (ROSA)26Sortm1(CAG‐S1pr1) transgene were crossed to Cdh5‐Cre ^ERT2 Cre males.

2. Strategy to induce expression of S1PR1 in post‐natal endothelium. Pups were given tamoxifen at P12 and P14, and retinas were analyzed at P17.

3. Flat‐mounted retinas from OIR S1pr1 ^f/stop/f and S1pr1 ECTG pups at P17. High‐magnification pictures of superficial capillaries showing S1PR1 induction in the S1pr1 ECTG.

4. Flat‐mounted retinas from OIR S1pr1 ^f/stop/f and S1pr1 ECTG pups at P17. Retinal vasculature parameters were stained with isolectin and images of superficial (top panel), intermediary (middle panel), and deep (lower panel) plexuses are shown.

5. Quantification of morphometric parameters in between retinas from S1pr1 ^f/stop/f and S1pr1 ECTG. Retinas from P9 S1pr1^f/stop/f mice were used as controls.

6. Dot plot representing expression level and frequency of cell types markers among non‐neuronal retinal cells at P14 and P17 in OIR.

7. Volcano plot showing expression level and frequency of OIR‐induced genes (left) and S1P‐related genes (right) among non‐neuronal retinal cells at P14 (left) and P17 (right) in OIR. S1pr4 and S1pr5 expression was limited to a low number of endothelial cells.

8. Cross sections from OIR S1pr1 ^f/stop/f , S1pr1 ECTG, and S1pr1 ECKO P17 pups stained for S1PR1 (red). Blood vessels are delineated by collagen IV (green) and nuclei are stained with Hoechst (blue) (left). Quantification of associated S1PR1 fluorescence inside blood vessels (right).

9. VEGF level in P17 retinas quantified by ELISA. Retinas from normoxic S1pr1 ^f/stop/f animal (n = 10) or OIR S1pr1 ^f/stop/f (n = 10) and S1pr1 ECTG pups (n = 13) at P17 were used.

10. Flat‐mounted retinas from OIR S1pr1 ^f/stop/f and S1pr1 ECTG pups at P15. Blood vessels are stained with isolectin (left). Avascular area, total neovascular tuft area, and average neovascular tuft size are quantified (right).

Data information: Data in (E, H and I) were analyzed by ANOVA test, and in (J) by one‐tailed Student's t‐test. Data are expressed as mean ± SD. A minimum of three pups per group were analyzed.

Figure 1. Endothelial S1PR1 signaling suppresses pathological neovascular tufts in OIR

1. Strategy to induce S1PR1 expression in the endothelium post‐OIR. S1pr1 ECTG pups were exposed to 75% oxygen from P7 to P12. Upon return to normoxia, S1PR1 expression was induced by tamoxifen injection at P12 and P14, and retinas were analyzed at P17.

2. Flat‐mounted retinas from OIR S1pr1 ^f/stop/f and S1pr1 ECTG pups at P17. Blood vessels are stained with isolectin (left) and pathological neovascular tuft area is highlighted in yellow. Avascular area and total and average neovascular tuft areas are quantified (right)

3. Flat‐mounted retinas from OIR S1pr1 ^f/f and S1pr1 ECKO pups, at P17. Blood vessels are stained with isolectin and neovascular tufts are highlighted in yellow. Avascular area and total and average neovascular tuft areas are quantified.

4. Cross sections from OIR S1pr1 ^f/stop/f , S1pr1 ECTG, and S1pr1 ECKO P17 pups. Blood vessels are stained with isolectin (red) and nuclei with Hoechst (blue).

Data information: Data in (B and C) were analyzed by one‐tailed Student's t‐test. Data are expressed as mean ± SD. A minimum of five pups were analyzed per group.

Figure 2. Endothelial S1PR1 enhances junctions while inhibiting vascular leakage and pericyte loss in OIR

1. Retinal flat mounts from OIR S1pr1 ^f/stop/f or S1pr1 ECTG or ECKO P17 pups. High‐magnification view of neovascular tufts stained for VE‐cadherin (green) and Claudin‐5 (red). Junctional levels of VE‐cadherin and Claudin‐5 were quantified as described, for a minimum of five animals per genotype.

2. Retinal flat mounts from OIR S1pr1 ^f/stop/f or S1pr1 ECTG or ECKO P17 pups. Sites of vascular leakage were assessed by staining for fibrinogen (red), and blood vessels are delineated by isolectin (green). Extravascular fibrinogen was quantified as described for a minimum of five animals per genotype.

3. Pericytes stained by NG2 on the surface of isolectin‐positive neovascular tufts at P17 post‐OIR (left). EC‐associated pericytes in neovascular tufts were quantified (right). A total of 70 tufts from three different animals were assessed.

4. Flat‐mounted retinas from OIR S1pr1 ^f/f and S1pr1 ECKO pups at P21 stained with isolectin. Avascular area and total and average neovascular tuft areas are quantified. A minimum of seven pups were analyzed per group.

Data information: Data in (A–C) were analyzed by ANOVA, and in (D) by one‐tailed Student's t‐test. Data are expressed as mean ± SD.

Figure EV2. Neovascular tuft phenotypes in S1pr1 ECTG at P17

1. Flat‐mount view of neovascular tufts from S1pr1 ^f/stop/f , S1pr1 ECTG, or ECKO at P17, stained for endothelial nuclei (ERG), blood vessel (isolectin), and associated endothelial nuclei quantification.

2. Flat‐mount view of neovascular tufts from S1pr1 ^f/stop/f or S1pr1 ECTG at P17, stained for immune cells (CD45), blood vessels (isolectin) (left), and FACS on retinal single‐cell preparations (right). Quantification of immune cells by FACS in OIR retinas from S1pr1 ^f/stop/f and S1pr1 ECTG at P17. Total immune cells (CD45 positive, left panel), monocytes/macrophages (CD11b positive, middle panel), and neutrophils (CD11b and Ly6G double positive, right panel) are presented.

3. Flat‐mounted retinas from OIR S1pr1 ^f/stop/f and S1pr1 ECTG pups at P19 stained with isolectin. Avascular area and total and average neovascular tuft areas are quantified.

Data information: Data in (A) were analyzed by ANOVA, and in (B, C) by one‐tailed Student's t‐test. Data are expressed as mean ± SD. A minimum of three pups per group were analyzed.

Figure 3. Systemic administration of S1PR1 agonists as a therapeutic strategy in retinopathy

1. Flat‐mounted retinas from OIR Apom ^+/− and Apom ^−/− pups at P17, stained with isolectin. Avascular area and total and average neovascular tuft areas are quantified (n = 10 pups/group).

2. Flat‐mounted retinas from OIR WT and Apom ^TG pups at P17, stained with isolectin. Avascular area and total and average neovascular tuft areas are quantified (n = 7 pups/group).

3. Flat‐mounted retinas from SAR247799‐treated, OIR WT pups at P17, stained with isolectin. Avascular area and total and average neovascular tuft areas are quantified (n = 10 pups/group).

4. Flat‐mounted retinas from ApoM‐Fc‐S1P‐treated, OIR S1pr1 ^f/stop/f , or S1pr1 ECTG pups at P17, stained with isolectin. Avascular area and total and average neovascular tuft areas are quantified (n = 6 pups/group).

Data information: Data in (A–C) were analyzed by one‐tailed Student's t‐test and in (D) by ANOVA. Data are expressed as mean ± SD.

Figure EV3. Effects of ApoM‐Fc‐S1P treatment on post‐OIR retinas

1. VEGF expression level in P17 retinas quantified by ELISA. Retinas from post‐OIR vehicle‐ or ApoM‐Fc‐S1P‐treated (n = 7 and 8, respectively) WT pups at P17.

2. Retinal flat mounts from OIR vehicle‐ or ApoM‐Fc‐S1P‐treated pups. High‐magnification view of neovascular tufts stained for VE‐cadherin (green) and Claudin‐5 (red). Junctional density was quantified as described by a minimum of seven animals per condition.

Data information: Data in (A and B) were analyzed by one‐tailed Student's t‐test. Data are expressed as mean ± SD.